Auto-Weighted Centralized Multi-Task Learning for Early MildCognitive Impairment Diagnosis
Cheng Nina1, Xiao Xiaohua2, Hu Huoyou2, Yang Peng1, Wang Tianfu1, Lei Baiying1*
1(Department of Biomedical Engineering, School of Medicine, Shenzhen University, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, Shenzhen 518060, Guangdong, China) 2(Department of Neurology, Shenzhen Second People′s Hospital, Shenzhen 518000, Guangdong, China)
摘要轻度认知障碍(MCI)是阿尔茨海默病(AD)的早期阶段,是治疗AD的最佳时期,因此对MCI的诊断非常重要。多模态数据可以全面分析疾病的状况,有利于疾病的准确诊断,但是现有方法并不能同时有效地分析多个模态数据之间的关系,无法有效结合功能态数据和结构态数据之间的优势。提出一种中心化自动加权多任务学习方法用于MCI的诊断。该方法可以同时学习不同模态的数据,有效地结合数据之间的优势。首先,分别对功能态数据rs-fMRI和结构态数据DTI构造脑网络;其次,基于多模态数据设计新的多任务特征学习模型,每个任务的重要性和模态之间的平衡关系会被自动学习,包括不同模态间的相似性和特异性,以获得稳定且有识别力的表达特征;最后,将选取的特征输入支持向量机模型进行分类诊断。实验基于Alzheimer′s Disease Neuroimaging Initiative(ADNI)公共数据库,包括明显记忆问题(SMC)、早期轻度认知障碍(EMCI)、晚期轻度认知障碍(LMCI)和正常受试者(NC)。所提出的方法对于NC vs SMC、SMC vs EMCI、SMC vs LMCI和EMCI vs LMCI等4种不同类型数据,诊断结果分别为76.67%、79.07%、80.56%和74.29%,与其他传统算法相比,分类准确率都有明显的提高,有望应用于对早期轻度认知障碍的诊断分析。
Abstract：Mild cognitive impairment (MCI) is the early stage of Alzheimer′s disease (AD) and is the best time for the diagnosis of the disease. By using multi-modal data one can comprehensively analyze the condition of the disease, which is conducive to the accurate diagnosis of the disease. However, the existing methods cannot effectively analyze the relationship between multiple modal data at the same time, and cannot effectively combine the advantages between functional state data and structural state data. For this reason, this paper proposed an auto-weighted centralized multi-task learning method for the diagnosis of MCI. The method can simultaneously learn data of different modalities and effectively combine the advantages between data. Specifically, firstly, brain networks were constructed for functional state data rs-fMRI and structural state data DTI respectively. Secondly, a new multi-task feature learning model was designed based on multi-modal data. The balance between the importance of each task and the modalities was automatically learned, including the similarity and specificity between different modalities, to obtain stable and discriminative expression features. Finally, the selected features were input into support vector machine (SVM) model for classification diagnosis. The experiments in this paper are based on the public database of the ADNI (Alzheimer′s Disease Neuroimaging Initiative), including significant memory concern (SMC), early mild cognitive Impairment (EMCI), late mild cognitive impairment (LMCI) and normal control (NC). The method had four different types of data for NC vs. SMC, SMC vs. EMCI, SMC vs. LMCI and EMCI vs. LMCI. The diagnostic result was 76.67%, 79.07%, 80.56% and 74.29%, respectively. Compared with other traditional algorithms, the classification accuracy of the methods described in this paper was significantly improved. The experimental results showed that the proposed method could be effectively applied to the diagnosis and analysis of early mild cognitive impairment.
成妮娜, 肖小华, 胡火有, 杨鹏, 汪天富, 雷柏英. 基于中心化自动加权多任务学习的早期轻度认知障碍诊断[J]. 中国生物医学工程学报, 2019, 38(6): 653-661.
Cheng Nina, Xiao Xiaohua, Hu Huoyou, Yang Peng, Wang Tianfu, Lei Baiying. Auto-Weighted Centralized Multi-Task Learning for Early MildCognitive Impairment Diagnosis. Chinese Journal of Biomedical Engineering, 2019, 38(6): 653-661.
 2018 Alzheimer′s disease facts and figures [J]. Alzheimers & Dementia, 2018, 14(3): 367-429.  Weiner W, Aisen S, Jack R, et al. The Alzheimer′s disease neuroimaging initiative: progress report and future plans [J]. Alzheimer′s & Dementia, 2010, 6(3): 202-211.  Podhorna J, Krahnke T, Shear M, et al. Alzheimer′s disease assessment scale-cognitive subscale variants in mild cognitive impairment and mild Alzheimer′s disease: change over time and the effect of enrichment strategies [J]. Alzheimer′s Research & Therapy, 2016, 8(1): 8.  Galea M, Woodward M. Mini-mental state examination (MMSE) [J]. Australian Journal of Physiotherapy, 2005, 51(3): 198-198.  Hua Xue, Leow D, Parikshak N, et al. Tensor-based morphometry as a neuroimaging biomarker for Alzheimer′s disease: an MRI study of 676 AD, MCI, and normal subjects [J]. Neuroimage, 2008, 43(3): 458-469.  Landau M, Harvey D, Madison M, et al. Associations between cognitive, functional, and FDG-PET measures of decline in AD and MCI [J]. Neurobiology of Aging, 2011, 32(7): 1207-1218.  Cuingnet R, Gerardin E, Tessieras J, et al. Automatic classification of patients with Alzheimer′s disease from structural MRI: A comparison of ten methods using the ADNI database [J]. Neuroimage, 2011, 56(2): 766-781.  Foster L, Heidebrink L, Clark M, et al. FDG-PET improves accuracy in distinguishing frontotemporal dementia and Alzheimer′s disease [J]. Brain, 2007, 130(10): 766-781.  Cherubini A, Péran P, Spoletini I, et al. Combined volumetry and DTI in subcortical structures of mild cognitive impairment and Alzheimer′s disease patients [J]. Journal of Alzheimer′s Disease, 2010, 19(4): 1273-1282.  Teipel J, Wohlert A, Metzger C, et al. Multicenter stability of resting state fMRI in the detection of Alzheimer′s disease and amnestic MCI [J]. Neuroimage: Clinical, 2017, 14: 183-194.  Hett K, Ta Vinh T, Catheline G, et al. Multimodal hippocampal subfield grading for Alzheimer′s disease classification [J]. Biorxiv, 2018, 29(3): 126-140.  Tong Tong, Gray K, Gao Qinquan, et al. Multi-modal classification of Alzheimer′s disease using nonlinear graph fusion [J]. Pattern Recognition, 2017, 63: 171-181.  Zhang Daoqiang, Wang Yaping, Zhou Luping, et al. Multimodal classification of Alzheimer′s disease and mild cognitive impairment [J]. Neuroimage, 2011, 55(3): 856-867.  Hinrichs C, Singh V, Xu Guofan, et al. Predictive markers for AD in a multi-modality framework: an analysis of MCI progression in the ADNI population [J]. Neuroimage, 2011, 55(2): 574-589.  Saeys Y, Inza I, Larrañaga P. A review of feature selection techniques in bioinformatics [J]. Bioinformatics, 2007, 23(19): 2507-2517.  Cui Zaixu, Zhong Suyu, Xu Pengfei, et al. PANDA: A pipeline toolbox for analyzing brain diffusion images [J]. Frontiers in Human Neuroscience, 2013, 7: 42.  Yan Chaogan, Zang Yufeng. DPARSF: a MATLAB toolbox for" pipeline" data analysis of resting-state fMRI [J]. Frontiers in Systems Neuroscience, 2010, 4: 13.  Craddock C, James A, Holtzheimer E, et al. A whole brain fMRI atlas generated via spatially constrained spectral clustering [J]. Human Brain Mapping, 2012, 33(8): 1914-1928.  Liu Guangcan, Lin Zhouchen, Yu Yong. Robust subspace segmentation by low-rank representation [C] //Proceedings of the 27th International Conference on Machine Learning (ICML-10). Brisbane: Springer, 2010: 663-670.  Tibshirani R. Regression shrinkage and selection via the lasso [J]. Journal of the Royal Statistical Society, 1996, 267-288.  Chen Jianhui, Zhou Jiayu, Ye Jieping. Integrating low-rank and group-sparse structures for robust multi-task learning [C] //Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, California: Springer, 2011: 42-50.  Benesty J, Chen Jingdong, Huang Yiteng, et al., On the importance of the pearson correlation coefficient in noise reduction [J]. IEEE Transactions on Audio Speech and Language Processing, 2008, 16(4): 757-765.  Hanley A, McNeil J. The meaning and use of the area under a receiver operating characteristic (ROC) curve [J]. Radiology, 1982, 143(1): 29-36.  Vasavada M, Wang Jianli, Eslinger J, et al. Olfactory cortex degeneration in Alzheimer′s disease and mild cognitive impairment [J]. Journal of Alzheimer′s Disease, 2015, 45(3): 947-958.  Gauthier S, Reisberg B, Zaudig M, et al. Mild cognitive impairment [J]. The Lancet, 2006, 367(9518): 1262-1270.  Gomar J, Bobes T, Conejero C, et al. Utility of combinations of biomarkers, cognitive markers, and risk factors to predict conversion from mild cognitive impairment to Alzheimer disease in patients in the Alzheimer′s disease neuroimaging initiative [J]. Archives of General Psychiatry, 2011, 68(9): 961-969.  张道强, 接标. 基于机器学习的脑网络分析方法及应用[J]. 数据采集和处理, 2015(1): 68-76.  Argyriou A, Evgeniou T, Pontil M. Multi-task feature learning [C] //Advances in Neural InformationProcessing Systems (NIPS 2007). Vancouver: Springer, 2007: 41-48.